The origin of star formation at z ≤ 1: secular or episodic ?

1. The origin of star formation at z ≤ 1: secular or episodic ? by François Hammer GEPI, Observatoire de Paris

2. A formation history with violent episodes revealed by panchromatic observations Hammer et al, 2005, A&A, 430, 115 (also astro-ph/0410518) GEPI, Observatoire de Paris Did most spirals form during the last 8 Gyrs ? Collaboration with: Hector Flores, Yan-Chun Liang, Xian-Zhong Zheng, François Assemat and David Elbaz (CEA/Sap) & Catherine Cesarsky (ESO)

3. 10.5 < logM/Mo< 11.5 Sa-Sbc E/S0 Irr A panchromatic survey of z=0.4-1 galaxies • 200 intermediate mass galaxies (3 1010 < Mstar < 3 1011MO, CFRS) • Locally, this mass range is made of (Nakamura et al, 2004, SDSS) : • 27% of E/S0, 53% of Sa-Sbc, 17% of Sc-Sd, 3% of Irr+other • from 65 to 75% of the global stellar mass (Brinchman & Ellis, 2000; Heavens et al, 2004)

4. Goals: “robust” evolutionary features from z=1 to z=0 • 200 intermediate mass galaxies (3 1010 < Mstar < 3 1011MO, CFRS) • Follow-up with VLT (15 nights), HST (400 orbits) & ISO (200 orb.), • by combining: • 1. High S/N spectroscopy & multi- photometry • Robust estimates of extinction, SFR & SFR/Mstar • “ “ of gas phase metal abundances • 2. High S/N HST images in 2 broad bands (F606W & F814W) • automatic software + colour maps + independent analyses • Simple classification scheme: E/S0, Sp, Irr, major mergers & objects too compact to be properly classified •  limit the possible selection effects (k corrections/dimming)

5. Methodology: references Liang et al, 2004a, “Misleading results from low resolution spectroscopy: from galaxy ISM chemistry to cosmic star formation density”, A&A, 417, 905 Flores et al, 2004, “Star formation rates of distant LIRGs derived from Ha and IR luminosities”, A&A 415, 885 Liang et al, 2004b, “The luminosity-metallicity relation of distant LIRGs”, A&A 423, 867 Zheng et al, 2004, “HST/WFPC2 morphologies and color maps of distant LIRGs”, A&A 421, 847 Zheng et al, 2005, “HST/WFPC2 morphologies and bar structures of field galaxies at 0.4 < z < 1”, A&A in press (astro-ph/0502170) Marcillac et al, 2005, “Deriving the star formation history of distant LIRGs from their optical spectra”, A&A in preparation

6. Galaxy spectroscopy pre-requisites(Liang et al, 2004, A&A 417, 905) • R>1000 spectroscopy for: • extinction ( Balmer lines corrected • for stellar absorption) • SFRs • gas chemistry • proper analysis of stellar populations

7. - extinction corrected H SFRs are close to mid-IR estimates (Elbaz et al, 2002) for SFR < 150 MO/yr (i.e. below ULIRGs) •  more robust SFR estimates • luminous IR galaxies (not ULIRGs) dominates the cosmic star formation density at z~1 • (confirmed by Spitzer, Le Floch et al, 2004) •  less than 20% of the star formation density is coming from extremely dust enshrouded regions Estimating extinctions and SFRs at z ~1(Flores et al, 2004, A&A 415, 885) FORS2/ISAAC: 16 ISO galaxies, 0.4< z <1

8. Extinctions and SFRs at z ~1 usingH/Hratio(Liang et al, 2004b, A&A 423, 867) • FORS2/R=1200: • 90 ISO galaxies at 0.5<z<1 • if properly corrected for underlying absorption: Av(H/H) consistent with Av(H/H) •  could be used until z=1 • - Extinction coefficient Av(IR) estimated from the energy balance between IR and H •  extinction/SFR estimate accuracies of ~ 0.6 mag at 5500A

9. SFRs as estimated by UV, [OII] & IR(Hammer et al, Venice 2003, proceedings, astro-ph/0401246) OII line & UV luminosities underestimate SFR values by factors 5 to 100 for starbursts & LIRGs !

10. LIRGs: potentially double their masses in ~0.8 Gyr SFR: IR & Ha Red dots: LIRGs (20-200 MO/yr) Full squares: starbursts (<20M/yr) SFR: [OII]3727 Open symbols From BE00: Brinchman & Ellis 2000

11. The impact of LIRGs to the SFD at z > 0.4 • At z=0: LIRGs represent ~ 0.5% of L > 1010LO (Soifer et al, 96) • 0.4 < z < 1: LIRGs represent 15% of intermediate mass galaxies • (confirmed by Spitzer, Yan et al, 2004) • The strongest & most evolving component in SF history (Elbaz & Cesarsky, 2003) • contribute to 50 to 66% of the SFR density at z= 0.7 LIRGs are intermediate mass galaxies : During the 3.3 Gyrs elapsed time (z=0.4-1), they have produced, 0.15 x 3.3/0.8 = 62% of the z=1 total mass of interm. mass galaxies (or 38% of their total mass at z=0)

12. How to account for the high LIRG fraction (15% of intermediate mass galaxies) ? • A specific population ? • LIRGs are continuously forming stars during 3.3 Gyrs (z=1  z=0.4) • they would multiply their masses by 2 x (3.3/0.8)=8.2 !! • BUT no trace of recent formation of massive galaxies, dominated by E/S0, with 3 1011<Mstar<31012MO

13. Episodic star formation history: IREs of galaxies • Most plausible: • most galaxies experienced IR • episodes (IREs) which • reddened them during • successive « short » periods • supported by spectral similarities between all types of the numerous emission line galaxies at z=0.4-1 (70%) • n= 5 (IRE/0.1Gyr)-1 per galaxy, or n= 1 IRE with IRE ~ 0.5 Gyr

14. SFRIR+UV SFRUV all “massive” systems were formed at z=1 ? challenged by • more robust estimates of stellar mass evolution: • 30% to 50% of the mass formed since z=1 (Dickinson et al, 2003) • It matches the stellar formation revealed in IR (2 to 3 times that detected at rest-frame UV, Flores et al, 1999) • mostly due to LIRGs (20 < SFR < 200MO/yr) • & analysis of fossil record of 105 SDSS galaxies show that intermediate mass galaxies have formed/assembled 60% of their mass since z=1 (Heavens et al, 2004) •  An update of the galaxy evolution scheme during the last 8 Gyrs

15. Metal-abundance relationship at 0.4< z < 1(Liang et al, 2004, A&A 423, 867) • At z = 0.4-1, intermediate mass galaxies show gas abundances (0.3 dex on average) lower than present day spirals • they can reach local disks locus assuming an episodic star formation history 03.0035 (from R23=(OII+OIII)/Hb)

16. Link to morphological changes ? Episodic star formation is consistent with hierarchical galaxy formation • Production of ~ 62% of the z=1 total mass (intermed. mass gal. ) • LIRGs are witnessing the major morphological changes • Merger (minor & major), gas infall related to galaxy interaction, cold gas flow ? LIRG morphologies

17. Morphological classification & V-I color maps for 31 LIRGs: 0.5< z< 1.1 5 1010 < LIR /LO < 5 1012 } Disks: 36% 03.0035 Zheng et al, 2004, A&A 421, 847 Color maps after V & I image alignement 0”.015 rms Irregulars: 22% Major mergers: 17% 03.1531 Compact: 25%

18. Major morphological changes in the general galaxy population 64% of LIRGs Morph. z ~ 0.7 (*) z=0 (**) Type E/S0 23% 27% Spiral 43% 70% ( 53% earlier than Sbc) Irregular 9% ~ 2% Major merger 6% < 2% Compact/LCG 19% < 2% Sources: (*) Zheng et al, 2004a (A&A 421, 847) & 2004b (**) Nakamura et al, 2004 (SDSS)

19. FORS2/R=1200 spectra : 30 LCGs: mix of strong emission lines & strong absorption lines Forming ~ 5-100 MO/yr Luminous compact galaxies (LCGs) at z ~ 0.7 (Hammer et al, 2001, ApJ, 550, 570 andGruel 2002, PhD thesis) Emerging population at z ~ 0.7 r50 < 3.5 h70-1 kpc & MB < -20 • As luminous as Milky Way within <2 times the size of LMC • stellar masses near M* • brighter end of Guzman ‘s LCBGs Suggested to be the formation of the bulge prior to the star formation in the disk (Hammer et al, 2001)

20. LIRGs/LCGs: related to which galaxy type ? z~0.7 spirals have blue cores  star formation in bulges z~0.7 spirals have blue cores  star formation in bulges • Production of ~ 40% of the z=0 total mass (intermediate mass galaxies) is (mainly?) occuring in spiral’s progenitors • It is made during violent & successive episodes (LIRGs) • related to major changes of galaxy morphology • LIRGs are major mergers (17%), LCGs (25%), Irr (22%) & Spirals (36%)  link those species together with a violent star formation history ?  If (LIRGs or LCGs  E/S0) over production of today E/S0 !

21. Ingredients for a spiral rebuilding scenario From z=1 to z~0.4: References • episodic SFH / morphological changes Hammer et al, 2004 • nmerger= 0.75 per galaxy Le Fèvre et al, 2000; Bundy et al, 2004 • LCGs: merger remnants Hammer et al, 2001 • Irregular: minor mergers • > 5 x more gas than in today spirals Pei, Fall & Hauser, 1999

22. A spiral rebuilding scenario Hydrodynamical models from Tissera et al (2002)

23. A spiral rebuilding scenario from Cox et al, 2004 Hydrodynamical models from Tissera et al (2002)

24. Reporting for the observed fractions Observed fraction Phase Characteristic time Theoretical time References (all spiral) (75% of spiral) [Gyr] [Gyr] [Gyr] Major merger* Compact** Irregulars Spirals 8% 24.5% 11.5% 56% 0.26 0.80 0.38 1.80 0.35 1.10 0.38 - 0.2-0.5 0.6-2.0 0.2-0.5 - (1) (1),(2),(3) (1) - (1) Tissera et al, (2000) (2) Baugh et al (1996) (3) Cox et al (2004)

25. Scenario Evolution Star formation only in dwarves (BE2000; Cowie et al, 1996) Spiral rebuilding (Our scenario) Minor merger induced starbursts (Somerville et al 2001) Heavens 2004 Dickinson 2004 Mass growth Hammer, 2004 Liang 2004 L-Z (O/H) Le Fèvre 2000 Bundy 2004 Pair statistics Flores 1999 Elbaz 1999 IR light evol. # density E/S0 Schade 1999 # density Sp Lilly 1998 # density of Pec. Brinchman, 1998 Ellis 2001 Zheng 2004 Sp core colours References

26. Scenario Object Galactic downsizing (BE2000; Cowie et al, 1996) Spiral rebuilding (Our scenario) Collisional starbursts (Somerville et al 2001) Prantzos 1995 Bouwens 1997 Milky Way Rich, 2004 Brown 2003 M31 References The stellar populations concept was developed based on Baade's studies of the central bulge of the Andromeda Galaxy, M31. Yet, the Andromeda system now appears to be anything but typical. Deep imaging reveals the outer halo to be a flattened system of complex tidal filaments roughly aligned with the disk major axis. Studies of the field remote from the nucleus find a metal-rich population that shows little variation in its mean abundance or distribution in projected locations ranging from 5 to 35 kpc from the nucleus. Every study to date confirms that the halo population is metal-rich, at [Fe/H]~ -0.5, with a tail extending to lower metallicity. The halo within 20 kpc cannot be composed of disrupted dwarf spheroidal satellites of M31, which are very metal-poor and have well-populated blue horizontal branches. The stars in the interaction streams of M31 are more metal-poor, with a smaller abundance spread than the disk stars.The age distribution of one deep halo field has been constrained from HST imaging, which shows that it has an old, metal-poor population, but also an equal population of 6-8 Gyr old, metal-rich stars. While the halo may be intermediate age, the bulge and a large fraction of the globular clusters appear to be old and similar to comparable populations in the Milky Way. We show that M31 lies in the expected location on a plot of halo abundances as a function of total galaxy luminosity. While this relationship may be understood as implying that a metal-enriched wind controls chemical evolution, it is difficult to understand how a simple closed-box model with wind outflow could be consistent with the presence of a significant age range in the halo. The complexity present in the old populations of M31 stands to remind us how complicated the history of distant galaxies might be.

27. Summary • Since z= 1 (last 8 Gyrs) : • 1- Intermediate mass spirals ~ doubled their mass/metal content • through episodic & violent bursts (IR episodes) 2- If 75+/-25% of spirals experiencing their last major merger event: •  simple link between distant and nearby galaxies; • consistent with simultaneous decreases of: star formation densities in UV & IR, merging rate, of number densities of LIRGs, compact, & Irr galaxies •  consistent with approximately constant densities of E/S0 and Sp 3- Stellar mass production: 42% in major mergers (collision & remnants) 22% in minor mergers (irregulars) 36% from gas infall (interactions ? cold gas ?)

28. An efficient way to form bulges in early type galaxies (75% of the present-day spiral population)

29. Is this simply crazy ??? Fusion of 2 spirals give an elliptical ? Yes if all the gas is exhausted! but see Springel & Hernquist (2004) Disks cannot support large SFRs ? but cold inflows can provide IREs (see Birboim & Dekel, 2004) Compact (LCGs) show narrow emission lines, so they should be associated with low mass galaxies (dwarf spheroidal, Guzman et al, 1997) ? ==> see velocity fields (Puech et al, 2005 and Ostlin et al 2001 for local) (Flores et al, 2004)